On Friday, July 12, 2002, at 02:06 , Connie Chan wrote:
[..] p0: I will defer to your understanding of chinese 'encryption' - since I haven't been in that since the wade-giles v. pin-yin debates.... but if it is a simple mapping as you suggest..... > b855=cdf2 > a456=d5c9 > a454=c8fd > a457=c9cf > a455=cfc2 > ...... where everything on the left hand side is 'uniq' the problem comes when you have say ff33 -> 1111 and ff33 -> 2222 but is context dependent on some ordering of characters before or after it.... So as long as they are always a simple one to one relationship you should be ok. > LHS is the Big5 Char, and RHS is GB2312. > So I tried to make a hash say > > $ch{'b855'} = 'cdf2' ; # something like that . > > Then I can operate it like this > $checkThis = 'b855'; > $newChar =~ s/($checkThis)/$ch{$1}/eg; # dunno it works or not, not tested > yet =) > > So... any warning here ? =) here is where I would argue for $new_char = ( exists $big_5{$check_this} ) ? $big_5{$check_this} : 'UNK'; since the possibility exists that there is a character in Big5 that is not in GB2312.... ciao drieux --- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]